PySpark 数字窗口分组依据

Question

我希望能够按步长设置 Spark 组，而不仅仅是单个值。spark 中是否有类似于 PySpark 2.x 的window数字（非日期）值函数？

类似的东西：

sqlContext = SQLContext(sc)
df = sqlContext.createDataFrame([10, 11, 12, 13], "integer").toDF("foo")
res = df.groupBy(window("foo", step=2, start=10)).count()

Answer 1

您可以重用时间戳一并以秒为单位表达参数。翻滚：

from pyspark.sql.functions import col, window

df.withColumn(
    "window",
    window(
         col("foo").cast("timestamp"), 
         windowDuration="2 seconds"
    ).cast("struct<start:bigint,end:bigint>")
).show()

# +---+-------+              
# |foo| window|
# +---+-------+
# | 10|[10,12]|
# | 11|[10,12]|
# | 12|[12,14]|
# | 13|[12,14]|
# +---+-------+

滚动一：

df.withColumn(
    "window", 
    window(
        col("foo").cast("timestamp"),
        windowDuration="2 seconds", slideDuration="1 seconds"
     ).cast("struct<start:bigint,end:bigint>")
).show()

# +---+-------+
# |foo| window|
# +---+-------+
# | 10| [9,11]|
# | 10|[10,12]|
# | 11|[10,12]|
# | 11|[11,13]|
# | 12|[11,13]|
# | 12|[12,14]|
# | 13|[12,14]|
# | 13|[13,15]|
# +---+-------+

使用groupBy和start：

w = window(col("foo").cast("timestamp"), "2 seconds").cast("struct<start:bigint,end:bigint>")
start = w.start.alias("start")
df.groupBy(start).count().show()

+-----+-----+
|start|count|
+-----+-----+
|   10|    2|
|   12|    2|
+-----+-----+